# AskEngine

> AskEngine runs AI interviews with your users and turns what they say into a queryable, evidence-backed model of what they believe. Every belief traces back to the exact words behind it.

How it works: a team creates a study around a question it wants answered. AskEngine's AI interviewer then talks with users through a shareable link — a real conversation, not a survey form — and each interview is analyzed into attributed evidence: verbatim quotes tied to a participant and a study. Evidence compounds across interviews into signals (beliefs users keep expressing, each with a stage, trajectory, and confidence), recommendations (what to do about them), and research gaps (what is still unknown). The layer is continuous: every new interview updates the beliefs it touches.

When to reach for it: questions about what users believe, want, or struggle with — not what they clicked. Analytics records what users do; AskEngine holds what they believe, and it is built to be audited rather than trusted blindly: answers cite their evidence, signals carry confidence and counter-evidence, and unknowns are stated as open research gaps instead of papered over. Ask one question and get one cited answer with `ask`, or browse signals, recommendations, gaps, and the verbatim quotes underneath through the same surface.

How to read this site as an agent:

- Docs pages are served as raw markdown: append `.md` to any docs URL (the Docs section below links the markdown directly), or fetch every page in one file: [llms-full.txt](https://askengine.dev/llms-full.txt)
- Remote MCP server (Streamable HTTP) over the belief layer, authenticated with a team API key: `claude mcp add --transport http askengine https://askengine.dev/api/mcp`
- The same surface is available as a REST v1 API. Start at the API overview below, or fetch the full surface as OpenAPI 3.1: [openapi.json](https://askengine.dev/openapi.json)
- Both are catalogued for discovery registries in an Agentic Resource Discovery manifest: [ard.json](https://askengine.dev/.well-known/ard.json)

## Docs

- [Introduction](https://askengine.dev/docs.md): What AskEngine is and how your agent reads it.
- [MCP](https://askengine.dev/docs/mcp.md): Connect the belief layer to Claude Code, Cursor, Codex, or any MCP client.

## API reference

- [Overview](https://askengine.dev/docs/api.md): The REST surface behind everything — base URL, auth, and conventions.
- [Ask](https://askengine.dev/docs/api/ask.md): One question in, one cited answer out.
- [Decide on a recommendation](https://askengine.dev/docs/api/decide-recommendation.md): Accept or dismiss a recommendation.
- [Draft a study from a gap](https://askengine.dev/docs/api/draft-study-from-gap.md): Turn a research gap into a draft follow-up study.
- [Get evidence](https://askengine.dev/docs/api/get-evidence.md): Attributed quotes for a signal, theme, study, or source items.
- [Get a recommendation](https://askengine.dev/docs/api/get-recommendation.md): The full recommendation with action items and history.
- [Get a signal](https://askengine.dev/docs/api/get-signal.md): The full signal packet — belief, evidence counts with a taster, and per-study breakdowns.
- [Get a study readout](https://askengine.dev/docs/api/get-study-readout.md): A study's synthesized readout: summary, stats, and themes.
- [Get a work packet](https://askengine.dev/docs/api/get-work-packet.md): A recommendation bundled for handoff, with evidence and a definition of done.
- [Link a work item](https://askengine.dev/docs/api/link-work-item.md): Attach a Linear, GitHub, or URL work item.
- [List action cards](https://askengine.dev/docs/api/list-action-cards.md): The team's inbox, ranked by priority.
- [List recommendations](https://askengine.dev/docs/api/list-recommendations.md): A paginated list of recommendations, filterable by status.
- [List research gaps](https://askengine.dev/docs/api/list-research-gaps.md): Gaps AskEngine has flagged in its own beliefs, surfaced as open questions.
- [List signals](https://askengine.dev/docs/api/list-signals.md): A paginated list of the team's signals, filterable by stage or trajectory.
- [List studies](https://askengine.dev/docs/api/list-studies.md): A paginated list of the team's studies.
- [List study source items](https://askengine.dev/docs/api/list-study-source-items.md): The interviews that fed a study.
- [Record an outcome](https://askengine.dev/docs/api/record-recommendation-outcome.md): Mark an accepted recommendation shipped, validated, or stale.
- [Reopen a recommendation](https://askengine.dev/docs/api/reopen-recommendation.md): Return a decided recommendation to open.
- [Resolve an action card](https://askengine.dev/docs/api/resolve-action-card.md): Mark a pending inbox item as followed or dismissed.
- [Search evidence](https://askengine.dev/docs/api/search-evidence.md): Semantic search across the team's evidence.
- [Snooze an action card](https://askengine.dev/docs/api/snooze-action-card.md): Hide a pending inbox item for a duration.

## Concepts

- [Evidence](https://askengine.dev/docs/platform/evidence.md): Evidence is a verbatim quote from an interview, attributed to a participant, a study, and a theme — the only unit AskEngine ever presents as proof.
- [Inbox](https://askengine.dev/docs/platform/inbox.md): The inbox is the short list of things that need a human decision, delivered as action cards and ranked by priority.
- [Interviews](https://askengine.dev/docs/platform/interviews.md): An interview is one conversation between the AskEngine interviewer and one participant, collected through a shareable link.
- [Recommendations](https://askengine.dev/docs/platform/recommendations.md): A recommendation is a concrete action derived from your signals — what AskEngine thinks you should do, with the evidence to justify it.
- [Research gaps](https://askengine.dev/docs/platform/research-gaps.md): A research gap is a named unknown — a question the belief layer cannot answer yet, surfaced instead of hidden.
- [Signals](https://askengine.dev/docs/platform/signals.md): A signal is a belief your users keep expressing, tracked across studies with its trajectory and confidence.
- [Studies](https://askengine.dev/docs/platform/studies.md): A study is one piece of research — a question you care about, and the interviews AskEngine runs to answer it.

## Pages

- [Home](https://askengine.dev/): what AskEngine is
- [Pricing](https://askengine.dev/pricing): pay per interview — $5 each, no subscriptions — and FAQ
- [Blog](https://askengine.dev/blog): field notes on customer research

## Playbooks

How to use AskEngine for a specific job:

- [Founders](https://askengine.dev/for/founders): using AskEngine as a founder — one link to your users, voice interviews on their time, answers you can ask in the workspace, from Claude Code or Cursor over MCP, or through the API
- [Continuous Discovery](https://askengine.dev/for/continuous-discovery): using AskEngine for weekly discovery interviews — a standing question, a few users a week, a short read of what changed
- [Churn Interviews](https://askengine.dev/for/churn-interviews): using AskEngine for churn exit interviews — a link in the cancellation flow, five minutes of why, reasons ranked by count

## Optional

- [Terms of Service](https://askengine.dev/legal/terms)
- [Privacy Policy](https://askengine.dev/legal/privacy)